2

Is it possible for engine A to be decisively better than engine B at short time controls (say, 5 min/game) but decisively weaker at long time controls (say, 1 hour/game)?

If the answer is yes, what could possibly be the source of the discrepancy? If the answer is no, why do we need long time control tournaments like the TCEC superfinal, when we could just get both engines to play each other at short time control much more often?

  • Do you have any particular engine in mind? – Qudit May 6 '19 at 8:07
  • No actually. I was wondering more about why we need to have long-time control tournaments such as the TCEC superfinal. – Allure May 6 '19 at 8:12
1

Why not? I don't have an example here, but cache implementation could be a major factor.

A cache algorithm that doesn't scale well with time will not perform well on a long time control. No matter how much memory you have, at longer control it will surely be filled up very easily and very quickly.

1

Here's an argument for how this might happen, but I don't know if it applies to real engines out there.

Imagine engines A and B are able to calculate to the same depth at short time controls, but due to better optimizations, B can calculate deeper at long time controls. And then imagine engine A has a better evaluation function. It seems to me that engine A might have an advantage at fast time control, but a disadvantage at slow time control.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.